Mamm Res DOI 10.1007/s13364-015-0232-2

ORIGINAL PAPER

Temperature influences the activity patterns of species in a large neotropical wetland

Thiago Bernardes Maccarini1 & Nina Attias1 & Ísis Meri Medri2 & Jader Marinho-Filho3 & Guilherme Mourão4

Received: 30 October 2014 /Accepted: 14 June 2015 # Research Institute, Polish Academy of Sciences, Białowieża, Poland 2015

Abstract In this study, we characterized the activity patterns short period in the middle of the morning and/or afternoon. and estimated the activity overlap between two armadillo spe- The proportion of activity overlap was 0.28 between the spe- cies ( novemcinctus and Euphractus sexcinctus)in cies, occurring primarily at the beginning of the night. At the Pantanal wetlands of Brazil. We report the effect of daily lower temperatures, yellow tended to leave their mean ambient temperature on activity onset and duration of burrows earlier and were active for longer periods. This is these armadillos. We captured seven yellow armadillos and probably an adaptation to their physiological constraints and three nine-banded armadillos and fitted them with very high limited thermoregulatory capabilities. Xenarthrans could be frequency radios and temperature loggers. By monitoring the used as models to obtain valuable information about how temperature of the environment (air temperatures and burrow physiology affects mammal behavior. temperatures), we inferred when the armadillo was active (above ground) or inactive (inside the burrow). Yellow arma- Keywords Thermoregulation . Euphractus sexcinctus . dillos were active mainly during the daytime and in the begin- Dasypus novemcinctus . Burrow . Niche partitioning . ning of the night, while nine-banded armadillos were active Physiological constraints mainly during the night. However, nine-banded armadillos presented a bimodal activity pattern, becoming active for a Introduction

Communicated by: Karol Zub The circadian activity period adopted by an is one of Electronic supplementary material The online version of this article the most effective and generalized ways to minimize the in- (doi:10.1007/s13364-015-0232-2) contains supplementary material, fluence of unfavorable biotic and abiotic factors (Layne and which is available to authorized users. Glover 1985). Circadian activity cycles allow an animal to anticipate environmental changes and choose the appropriate * Nina Attias [email protected] moment for a specific response or activity (Aronson et al. 1993). Most armadillos (Mammalia: ) have a strategy for 1 Ecology and Conservation Graduate Program, Centro de Ciências avoiding unfavorable factors: they build or use burrows. Biológicas e da Saúde - Cidade Universitária, Federal University of Mato Grosso do Sul, s/n - Caixa Postal 549, Campo Grande, MS, Burrows can be used not only to escape from predators or Brazil for sleeping (McDonough and Loughry 2003; Desbiez and 2 Ecology Graduate Program, Centro de Ciências Biológicas e da Kluyber 2013) but also for thermoregulation (González et al. Saúde - Cidade Universitária, University of Brasilia, s/n - Caixa 2001). Burrows work as temperature buffers (González et al. Postal 549, Campo Grande, MS, Brazil 2001); armadillos can exit them once air temperatures are 3 Zoology Department, University of Brasilia, Campus Universitário favorable and return to them to maximize their energy Darcy Ribeiro, Brasília, DF, Brazil budgets. 4 Embrapa Pantanal, Brazilian Agricultural Research Corporation, Rua Mammalian circadian rhythms are responsive to environ- 21 de Setembro 1880, Corumbá, MS, Brazil mental factors, especially light cycles (Aronson et al. 1993). Mamm Res

However, armadillos are imperfect homeotherms and have McDonough 2013). Hence, the available ecological informa- limited thermoregulatory capabilities, low body temperatures, tion might not reflect patterns that can change from place to and low basal rates of metabolism (McNab 1985). These char- place, such as behavior. acteristics are probably associated with their burrowing activ- Therefore, the purpose of this study was to characterize the ity and feeding habits (McNab 1979, 1980). As a result of this activity patterns of these two species of armadillo in the unique characteristic among , air temperature Brazilian Pantanal, estimate their activity overlap, and deter- strongly affects the activity patterns of armadillos and other mine the influence of air temperature on their activity patterns. xenarthrans (Layne and Glover 1985; Chiarello 1998;Mourão and Medri 2007). The influence of temperature on the activity patterns of these can be determined from their activity Methods duration and timing (activity onset). Temporal niche partitioning is a mechanism that related Study area species adopt to avoid competition: different circadian activity patterns might result in differentiation of the use of resources The Pantanal is a large wetland region covering approximately or levels of predation susceptibility. It can facilitate coexis- 210,000 km2 of seasonal floodplain of Rio Paraguay and in- tence through avoidance of direct confrontation (interference cludes areas of Brazil, Bolivia, and Paraguay (Mittermeier competition) or through the reduction of overlap in resource et al. 2003). The Brazilian portion of Pantanal is located in demand (resource competition; Kronfeld-Schor and Dayan the western part of the country, near the geographic center of 2003). Partitioning of the temporal dimension of the niche South America, at about 100 m in elevation. We conducted between two species implies a smaller niche overlap than ex- our study at the Nhumirim Ranch (18°59′ S, 56°39′ W), an pected by chance. In this phenomenon, the less competitive experimental station of approximately 43 km2 located in the species, or subordinates, usually modify their activity to sub- Nhecolândia subregion in the state of Mato Grosso do Sul, optimal periods of the circadian cycle, reducing competition Brazil. The Nhecolândia subregion is characterized by a land- with the dominant species (Gutman and Dayan 2005). scape mosaic of savannas, scrub savannas, forests, and per- However, when competition is not the most important factor, manent and temporary ponds (Alho et al. 1987). the activity periods of similar species can tolerate an overlap The climate of the region is tropical semi-humid, and the higher than that expected by chance. This pattern might indi- average annual temperature is 25.5 °C (Calheiros and Fonseca cate that the activity periods of both species are influenced by 1996). Summers are hot and rainy, averaging 1100 mm rain- similar responses to external factors (e.g., external temperature fall and maximum air temperatures exceeding 40 °C. Winters or potential predators). Another hypothesis is that phylogenet- are warm and dry, averaging less than 300 mm rainfall and air ic constraints of the activity period make closely related spe- temperatures of 20 °C, though occasional cold fronts from the cies present similar activity patterns, limiting the opportunities south can cause air temperatures as low as 0 °C (Soriano et al. for temporal niche partitioning among closely related compet- 1997). itors (Roll et al. 2006). Even though there is no evidence to confirm this fact, this could be the case for armadillos given Data collection that their physiology makes them very sensitive to tempera- ture variations. During September 2007 and February–May 2010, we The yellow armadillo, Euphractus sexcinctus (Linnaeus searched for armadillos by foot or in an all-terrain vehicle, 1758), and the nine-banded armadillo, Dasypus novemcinctus catching eight armadillos by hand and two in live traps from (Linnaeus 1758), are sympatric burrowing armadillos that simultaneous studies focused on medium-sized carnivores in have overlapping dietary habits (Anacleto 2007) and may oc- the area. Of these, seven were yellow armadillos cupy the same habitat type in the Brazilian Pantanal. The (E. sexcinctus) and three were nine-banded armadillos former is a diurnal and conspicuous species, widely distribut- (D. novemcinctus). We placed them in a safe and ventilated ed throughout Brazil (except Amazonia) and South America, box and then transported them to the ranch laboratory. They and is very common in the Pantanal (Eisenberg and Redford were anesthetized there using an intramuscular dose of 1999). Despite these characteristics, its ecology is poorly 4 mg/kg Zoletil® (1:1 tiletamine hydrochloride and known, and studies on its activity patterns are rare (Bonato zolazepam hydrochloride; Virbac S.A., Cedex, France). et al. 2008). The latter is a predominantly nocturnal species They were weighed, measured, and marked with numbered (Emmons and Feer 1997; Eisenberg and Redford 1999)thatis ear tags (model 1005-4; National Band & Tag Co., Newport, found from Argentina to the USA, which is considered as the KY, USA). largest geographic range among armadillos (Wetzel 1985). With the animals under anesthesia, we attached a package Even though this is the most studied species, much of the (≈35 g) containing a StowAway Tidbit temperature data log- research has been done in the USA (Loughry and ger (Onset Computer Corporation, Bourne, MA, USA; Mamm Res accuracy, ± 0.4 °C at 20 °C) combined with a very high fre- whether our method was reliable at inferring when the arma- quency radio transmitter (frequency range 164–166 MHz; dillos were inside or outside the burrows, we also performed model R2030; Advanced Telemetry Systems, Isanti, MN, an armadillo activity simulation (Appendix 1). USA) to each armadillo. Each transmitter was equipped with We tracked the transmitter/logger packages using a hand- a mortality signal switch adjusted for 24 h. held directional antenna (RA-2AK; Telonics, Mesa, AZ, To find the most durable and safe way to attach the pack- USA) and a portable receiver (Telonics TR-4). Nine out of ages, several techniques were used. Some devices were at- ten packages naturally fell off the armadillos in the course of tached to the armadillos’ tails using rubber strips covered with their semi-fossorial habits, after which they were recovered adhesive tape or only layers of adhesive tape. Others were from the ground or burrows. attached to the edge of the carapace, at the right posterior of the body, by drilling holes in the carapace and passing cable Analysis ties through them. To perform the latter, we injected the ani- mals with a local dose of xylocaine (Eurofarma Lab., São Temperature data from all loggers were downloaded using the Paulo, Brazil). On the same day, once the anesthesia effects software Boxcar® Pro 4.0 (Onset®). To infer when the arma- subsided and all procedures were completed, we released the dillos were active (above ground) or inside the burrow, we armadillos at the same location where they were captured. compared Tarmadillo with Tburrow and Tair simultaneously. The temperature registered by the data logger attached to Whenever Tarmadillo was more stable and approximately equal the armadillo was termed Tarmadillo and was affected by the to Tburrow, we inferred that the animal was inside the burrow. If instantaneous ambient temperature and the heat emitted by instead it was approximately equal to Tair, we inferred that the the armadillo’s body. However, because body temperature animal was active and outside the burrow (Appendix 1). We varies according to the ambient temperature, but with less used the values for sunrise and sunset provided by the US variation (Roig 1969), most of the fluctuations in Tarmadillo Department of Commerce, National Oceanic and was attributed to the ambient temperature of the location of Atmospheric Administration (http://www.esrl.noaa.gov/)to the armadillo at a given moment. To obtain temperatures of define daytime and nighttime periods for the days of the burrow (Tburrow)andair(Tair), we placed a temperature monitoring. Daytime was defined as the period between logger inside an armadillo burrow (≈1 m inside) and another sunrise and sunset and nighttime as the remaining hours of in the meteorological station of the Nhumirim ranch. For each the day. To account for photoperiod variation, we corrected monitored animal, Tburrow was obtained by monitoring one the time of each activity record according to the photoperiod active burrow selected in the surroundings of the capture lo- duration of each day. cation. We cannot ensure that the selected burrow was the one To characterize the activity period of each species and the in use by the captured armadillo at that time. Nonetheless, our degree of overlap in the activity periods between species, we intention was not to record the temperature of any specific used a continuous, non-parametric model of a conditional, burrow as the armadillo could use more than one burrow dur- circular kernel function (Oliveira-Santos et al. 2013). Kernel ing the monitoring period. By placing the logger inside an density functions provide a continuous measure of the density active burrow our intention was to verify the temperature of data points throughout their distribution (Worton 1989). buffering effect of burrows. All temperature loggers were set Much like its use to calculate home ranges, this function al- to register the temperature every 5 min. When necessary, we lows us to consider the use of time in a continuous probabi- obtained the air temperature hourly from the automatic mete- listic way, providing a clearer measure of the probability of an orological station. animal being active at a given time within its activity period.

Whenever Tarmadillo was more stable and approximately A two-sample Kolmogorov–Smirnov test was used to test parallel to Tburrow, we inferred that the animal was inside the whether the distribution of activity records as a function of air burrow and inactive. Whenever Tarmadillo presented more pro- temperature (1 °C interval) differed from the distribution of all nounced variations following a Tair variation pattern, we in- temperature records registered during the study period (i.e., its ferred that the animal was active and outside the burrow. It Bavailability^). We used graphical analyses to evaluate how was possible to identify the activity onset by the abrupt vari- the armadillos' activity patterns were related to Tair and to ation in Tarmadillo values, after a sequence of similar values verify whether activity onset and duration were affected by with smother variations (see details in Appendix 1). To verify the mean Tair. Activity onset was defined as the time an arma- the accuracy of our method, we visually checked the activity dillo left its burrow each day. In cases where armadillos left status of the animals (inside or outside the burrow) once or their burrows more than one time per day, usually involving twice a day by locating the animal through radio telemetry. We one or more short periods and one long activity period, we performed at least nine observations per animal during their considered only the longest period as the onset time of activ- monitoring period and compared their observed activity status ity. In order to avoid pseudo-replication, we calculated the with the status inferred from the temperature data. To assess average Tair during the monitoring period for each animal Mamm Res and associated it with their average activity onset and dura- tion. Spearman correlation test was used to test if average Tair affected the mean time of day that yellow armadillos exited their burrows and their mean activity durations. Spearman correlation test was also used to evaluate the relation of pho- toperiod duration during each animal's monitoring period and the onset and duration of activity periods. Because we had data on only three nine-banded armadillos, we did not attempt to perform the correlation tests for this species, but we includ- ed the data on the graphs. All analyses were performed using R (R Development Core Team 2013).

Results

Of the seven captured yellow armadillos, two were male and five were female; of the nine-banded armadillos, one was male and two were female. Data logger packages remained on the Fig. 1 Distribution of temperatures registered during the monitoring armadillos for a period of 2 to 13 days (Supplementary mate- period at the Nhumirim ranch, Nhecolândia Pantanal, Mato Grosso do — Sul state, Brazil (September 2007 and February–May 2010). Solid bars rial Appendix 1). The longest duration was achieved by indicate the relative frequency at which each air temperature was using rubber strips covered with adhesive tape. However, registered during the monitoring period (ndn=12,473; nes =14,541). one yellow armadillo lost part of its tail after the package Dashed bars indicate the relative frequency of each air temperature was attached using this technique. It could have been affixed during active periods of a nine-banded armadillo (D. novemcinctus, n= 2813) and b yellow armadillo (E. sexcinctus, n=1822) too tight to allow normal blood circulation. Fortunately, the wound closed up well, and the armadillo was recaptured as beginning of the night (1806 to 0048 hours), before sunrise well as photographed several times by camera traps months (0324 to 0515 hours), and in the middle of the morning (0830 later. After this experience, we avoided using rubber strips and to 0900 hours). Among the yellow armadillos, mean T af- used only adhesive tape for the attachment of transmitters. air fected the mean activity onset positively (rho=0.82, p=0.03; The amplitude of recorded T was always smaller armadillo Fig. 3) and mean activity duration negatively (rho=−0.96, than that of T (Fig. 1). T was far more stable (range= air burrow p <0.01; Fig. 4). There was no correlation between 23.6–30.9 °C; CV=0.07) than Tair (range=5.7–40 °C; CV= 0.29), and the differences between them (Tburrow−Tair)ranged from −11.6 °C to +18.3 °C. This pattern allowed a confident estimate of the activity period of the monitored animals and the efficiency of the proposed method (for method details, see Supplementary material—Appendix 2). Nine-banded arma- dillos were active in a narrower and milder range of Tair (range=13.4–29.3 °C; CV=0.14) compared to yellow arma- dillos (range=13.2–36.2 °C; CV=0.20; Fig. 1). The distribu- tion of activity records as a function of air temperature differed from the distribution of all temperature records registered dur- ingthestudyperiod(forD. novemcinctus, D=0.2678, p<0.001; for E. sexcinctus, D=0.1587; p<0.001). The yellow armadillos were active for an average of 4.42 h/ day, mainly during the day and at the beginning of the night. The nine-banded armadillos were active for an average of 4.97 h/day and had primarily nocturnal activity. However, they presented a bimodal activity pattern, leaving the burrows for short periods in the middle of the morning and/or in the middle of the afternoon (Fig. 2). The proportion of activity overlap between the two species was 0.28 (smoothing param- Fig. 2 Kernel activity density estimates for nine-banded armadillo (D. novemcinctus, n=3; solid line) and yellow armadillo (E. sexcinctus, eter=37.4; isopleth=0.95), which is considered as a low over- n=7; dashed line) during a 24-h period. Gray areas represent periods of lap (Oliveira-Santos et al. 2013). The overlap occurred at the higher activity overlap (95 % kernel) between the species Mamm Res

during the first half of the night, irrespective of air temperature. Our data show that, as daily mean temperature decreases, yellow armadillos tend to leave their burrows earlier and be active for longer periods, maximizing their diurnal activity (Figs. 3 and 4). The zone of thermal comfort for yellow arma- dillos is in air temperatures of around 30 ° C. When subjected to ambient temperatures lower or higher than 30 ° C, the yellow armadillos show signs of excitation and more variable body temperatures (Roig 1969). Hence, yellow armadillos seem to be physiologically adapted to warm air temperatures (~30 ° C). Yellow armadillos presented a relatively higher proportion of activity records at a certain air temperature range (Fig. 1) and changes in activity behavior according to air temperature variation (Figs. 3 and 4). These could be evidences that the monitored animals are able, to a certain point, to adjust their Fig. 3 Response of the armadillos’ mean activity onset to the mean air behavior to deal with their physiological constraints that re- temperature. Black dots represent yellow armadillos (E. sexcinctus, n=7). quires them to be active at relatively warm air temperatures. Gray dots represent the nine-banded armadillo (D. novemcinctus, n=3) This is also supported by empirical observations in the Pantanal wetlands, where yellow armadillos are commonly photoperiod duration and activity onset (rho=0.32, p=0.40) seen active during the hottest periods of the day when most − or duration (rho= 0.64, p=0.1). animals are sheltered. It is well established that some species of armadillo can adapt their activity patterns in response to changes in temper- Discussion ature (McNab 1980; Breece and Dusi 1985; McDonough and Loughry 2003; Superina and Boily 2007). For example, the Despite our small sample size, it was possible to conclude that nine-banded armadillo and the little hairy armadillo air temperature affected the activity onset and duration of the ( vellerosus) tend to be more diurnal in cold studied yellow armadillos. We did not have enough data to winters and more nocturnal in the summers at higher temper- perform this analysis with nine-banded armadillos, but the atures (Layne and Glover 1985;Greegor1985). In tropical three monitored individuals generally exited their burrows areas, the yellow armadillo is generally classified as diurnal (Emmons and Feer 1997; Eisenberg and Redford 1999; Trolle 2003; Cuéllar 2008), while the nine-banded armadillo is re- ported as nocturnal (Eisenberg and Redford 1999;Trolle 2003). Our data suggest that, in similar air temperatures, the nine- banded armadillo tends to leave its burrow later than the yel- low armadillo (Fig. 3). Nonetheless, the yellow armadillo was both diurnal and nocturnal, as previously reported by Bonato et al. (2008). The nine-banded armadillo tended to be noctur- nal, but there were short periods of diurnal activity that con- firm an earlier report by Layne and Glover (1985). These activity patterns indicated a low activity overlap (0.28) be- tween these armadillo species. This could be considered as a preliminary evidence of separation in the temporal niches of the two species. However, this separation could be due to a variety of causes, and further studies should be performed to clarify this matter. Several studies focused on thermoregulation strategies of anteaters (Montgomery and Lubin 1977; Shaw et al. 1987; Fig. 4 Response of the armadillos’ mean activity duration to mean air temperature. Black dots represent yellow armadillos (E. sexcinctus, n=7). Medri and Mourão 2005; Mourão and Medri 2007), sloths Gray dots represent nine-banded armadillos (D. novemcinctus, n=3) (Chiarello 1998; Urbani and Bosque 2007) and armadillos Mamm Res

(Roig 1969; McNab 1980; Superina and Boily 2007). 2006; Mourão and Medri 2007); thus, the nine-banded arma- However, we could not find studies that focused on the activ- dillo is somewhat sheltered from large temperature variations ity onset of xenarthrans, with the exception of Camilo-Alves and extremes. Conversely, the yellow armadillo needs to and Mourão (2006). Their results indicated that the air tem- avoid extreme air temperatures, frequently experienced in perature linearly and positively affected the activity onset of open habitats, by returning to its burrow to Bregulate^ its body giant anteaters. The activity onset of the yellow armadillos temperature. Additionally, yellow armadillos are omnivorous seems to be even more affected by air temperature than that and have a broader diet compared to the more insectivorous of the (Camilo-Alves and Mourão 2006). diet of nine-banded armadillos (Anacleto 2007); hence, they Factors limiting temporal niche switches may be divided are not as influenced by the availability of invertebrate prey. into internal (i.e., anatomic or physiological, such as sensory Studies concerning the activity budgets of armadillos are and thermoregulatory capabilities) and external (i.e., biotic important because they may be conspicuously affected by and abiotic environmental conditions; Hut et al. 2012). their physiological constraints, such as low metabolic rate Because armadillos are imperfect homeotherms (McNab and high thermal conductance of the carapace, interacting 1985) and are primarily guided by odor and tactile stimuli, with large-scale environmental conditions, such as the climate with very poor eyesight (Emerling and Springer 2014), it is or landscape. In addition, physiological aspects may be more not at all surprising that their activity pattern is mainly regu- influential, affecting their behavior and activity more than lated by temperature rather than light. Their foraging activity other ecological traits, such as food availability. Armadillos does not seem to be harmed by this adaptive behavior, and and other xenarthrans could be used as models to obtain valu- they must gain physiological benefits. able information about how physiology affects decision mak- Although the variability of the activity period durations of ing in mammals. both armadillo species was high, the mean activity durations of the yellow armadillo and the nine-banded armadillo were similar and short. Armadillos are known to be active above- Acknowledgments This study was performed under License No. 006/ 2007, granted by the Brazilian Institute of Environment and Renewable ground a few hours per day (McDonough and Loughry 1997) Natural Resources, and License No. 22906-1, granted by the Chico and spend the rest of their time inside the burrow, either inac- Mendes Institute for Biodiversity Conservation. All applicable interna- tive or performing activities such as nesting, sheltering, or tional and national guidelines for the care and use of animals were follow- thermoregulating (González et al. 2001; McDonough and ed. We thank William J. Loughry for useful comments on the manuscript draft and all colleagues and friends who helped with the fieldwork and the Loughry 2003; McDonough and Loughry 2008). Such a short manuscript. We thank the two anonymous referees for their valuable period of activity is frequently associated with body tempera- comments and suggestions. We are grateful to Embrapa Pantanal for ture regulation (González et al. 2001). Since xenarthrans have logistic support, Program of Long Term Ecological Research/Conselho a low basal metabolic rate, about 40–60 % of what would be Nacional de Desenvolvimento Científico e Tecnológico (520056/98-1) for financial support, IDEA WILD for donating equipment, Conselho expected for a placental mammal of similar weight (McNab Nacional de Desenvolvimento Científico e Tecnológico for the produc- 1985), they have a limited capability to control body temper- tivity fellowship awarded to G. Mourão and J. Marinho-Filho and the ature and use burrows as thermal shelters. scholarships to Í.M. Medri and T.B. Maccarini, and Coordenação de McDonough and Loughry (1997) hypothesized that the Aperfeiçoamento de Pessoal de Nível Superior for the scholarships awarded to N. Attias and Í.M. Medri. levels of activity of nine-banded armadillos might increase on nights with higher soil temperatures and lower humidity. In these conditions, insects and other invertebrates, some of the main food sources for armadillos (Breece and Dusi 1985; References McDonough and Loughry 2008), would be less available, and thus armadillos would have to spend more time active above- Alho CJR, Lacher TE Jr, Campos ZMS, Gonçalves HC (1987) Mamíferos da fazenda nhumirim, sub-região de nhecolândia, ground to be able to find sufficient food. We were not able to pantanal do mato grosso do Sul: I – levantamento preliminar de determine if the activity period of nine-banded armadillos was espécies. Rev Bras Zool 4:151–164 affected by the air temperature in the Pantanal because our Anacleto TCS (2007) Food habits of four armadillo species in the Cerrado sample size was small. However, the periods of activity of Area, Mato Grosso, Brazil. Zool Stud 46:529–537 yellow armadillos tended to be negatively related to average Aronson BD, Bell-Pedersen D, Block BD, Bos NPA, Dunlap JC, Eskin A, Garceau NY, Geusz ME, Johnson KA, Khalsa SBS, Kosern-Van daily temperature, disagreeing with what was expected by the Hoffen GC, Koumenis C, Lee TM, LeSauter J, Lindgren KM, Liu McDonough and Loughry (1997) hypothesis. Such results Q, Loros JJ, Michel SH, Mirmiran M, Moore RY, Ruby NF, Silver might be related to the fact that, in contrast to the nine- R, Turek FW, Zatz M, Zucker I (1993) Circadian rhythms. Brain Res banded armadillo, the yellow armadillo spends most of its Rev 18:315–333 time in open areas, such as and forest edges Bonato V, Martins EG, Machado G, da Silva CQ, dos Reis SF (2008) Ecology of the armadillos unicinctus and Euphractus (Eisenberg and Redford 1999; Trolle 2003). Forested habitats sexcinctus (Cingulata: ) in a Brazilian Cerrado. J work as temperature buffers (Camilo-Alves and Mourão Mammal 89:168–174 Mamm Res

Breece GA, Dusi JL (1985) Food habits and home ranges of the common McDonough CM, Loughry WJ (2008) Behavioral ecology of armadillos. long-nosed armadillo Dasypus novemcinctus in Alabama. In: In: Vizcaíno SF, Loughry WJ (eds) The biology of the . Montgomery GG (ed) The evolution and ecology of armadillos, University Press of Florida, Gainesville, pp 281–293 sloths and vermilinguas. Smithsonian Institution Press, McNab BK (1979) The influence of body size on the energetic and dis- Washington, pp 419–427 tribution of fossorial and burrowing mammals. Ecology 60:1010– Calheiros DF, Fonseca Jr. WC (1996) Perspectivas de estudos ecológicos 1021 sobre o Pantanal. Embrapa CPAP Documentos. Document no. 18 McNab BK (1980) Energetics and the limits to a temperate distribution in Camilo-Alves CSP, Mourão GM (2006) Responses of a specialized in- armadillos. J Mammal 61:606–627 sectivorous mammal (Myrmecophaga tridactyla) to variation in am- McNab BK (1985) Energetics, population biology, and distribution of bient temperature. Biotropica 38:52–56 xenarthrans, living and extinct. In: Montgomery GG (ed) The evo- Chiarello AG (1998) Activity budgets and ranging patterns of the Atlantic lution and ecology of armadillos, sloths and vermilinguas. forest maned sloth Bradypus torquatus (Xenarthra: Bradypodidae). Smithsonian Institution Press, Washington, pp 219–232 J Zool (Lond) 246:1–10 Medri ÍM, Mourão G (2005) A brief note on the sleeping habits of the Cuéllar E (2008) Biology and ecology of armadillos in the Bolivian giant anteater—Myrmecophaga tridactyla Linnaeus (Xenarthra, Chaco. In: Vizcaíno SF, Loughry WJ (eds) The biology of the Myrmecophagidae). Rev Bras Zool 22:1213–1215 Xenarthra. University Press of Florida, Gainesville, pp 306–312 Mittermeier RA, Mittermeier CG, Brooks TM, Pilgrim JD, Konstant WR, Desbiez ALJ, Kluyber D (2013) The role of giant armadillos (Priodontes da Fonseca GAB, Kormos C (2003) Wilderness and biodiversity – maximus) as physical ecosystem engineers. Biotropica 45:537–540 conservation. Proc Natl Acad Sci U S A 100:10309 10313 Eisenberg JF, Redford KH (1999) Mammals of the Neotropics—the cen- Montgomery GG, Lubin YD (1977) Prey influences on movements of tral Neotropics: Ecuador, Peru, Bolivia. University of Chicago neotropical anteaters. In: Phillips RL, Jonkel C (eds) Proceedings of Press, Chicago the 1975 predator symposium. University of Montana, Missoula, pp – Emerling CA, Springer MS (2014) Genomic evidence for rod 103 131 monochromacy in sloths and armadillos suggests early subterranean Mourão G, Medri IM (2007) Activity of a specialized insectivorous mam- history for Xenarthra. Proc R Soc B. doi:10.1098/rspb.2014.2192 mal (Myrmecophoga tridactyla) in the Pantanal of Brazil. J Zool (Lond) 271:187–192 Emmons L, Feer F (1997) Neotropical rainforest mammals: a field guide, Oliveira-Santos LGR, Zucco CA, Agostinelli C (2013) Using conditional 2nd edn. University of Chicago Press, Chicago circular kernel density functions to test hypotheses on animal circa- González EM, Soutullo A, Altuna CA (2001) The burrow of Dasypus dian activity. Anim Behav 85:269–280 hybridus (Cingulata: Dasypodidae). Acta Theriol 46:53–59 R Development Core Team (2013) R: A language and environment for Greegor DH Jr (1985) Ecology of the little hairy armadillo statistical computing. R Found Stat Comp Roig VG (1969) Chaetophractus vellerosus. In: Montgomery GG (ed) The evolution Termorregulación en Euphractus sexcinctus (Mammalia: and ecology of armadillos, sloths and vermilinguas. Smithsonian – – Dasypodidae). Physis 29:27 32 Institution Press, Washington, pp 397 405 Roll U, Dayan T, Kronfeld-Schor N (2006) On the role of phylogeny in Gutman R, Dayan T (2005) Temporal partitioning: an experiment with determining activity patterns of rodents. Evol Ecol 20:479–490 – two species of spiny mice. Ecology 86:164 173 Shaw JH, Machado-Neto J, Carter TS (1987) Behavior of free-living Hut RA, Kronfeld-Schor N, van der Vinne V, Iglesia H (2012) In search giant anteaters (Myrmecophaga trydactyla). Biotropica 19:255–259 of a temporal niche: environmental factors. Prog Brain Res 199: Soriano BMA, Oliveira H, Catto JB, Comastri Filho JA, Galdino S, Salis – 281 304 SM (1997) Plano de utilização da fazenda Nhumirim. Embrapa Kronfeld-Schor N, Dayan T (2003) Partitioning of time as an ecological CPAP Documentos. Document No. 21 resource. Annu Rev Ecol Evol Syst 34:153–181 Superina M, Boily P (2007) Hibernation and daily torpor in an armadillo, Layne JN, Glover D (1985) Activity patterns of the common long-nosed the pichi (Zaedyus pichiy). Comp Biochem Physiol A 148:893–898 armadillo Dasypus novemcinctus in south-central Florida. In: Trolle M (2003) Mammal survey in the Southeastern Pantanal, Brazil. Montgomery GG (ed) The evolution and ecology of armadillos, Biodivers Conserv 12:823–836 sloths and vermilinguas. Smithsonian Institution Press, Urbani B, Bosque C (2007) Feeding ecology and postural behavior of the Washington, pp 407–417 three-toed sloth (Bradypus variegatus flaccidus) in northern Loughry WJ, McDonough CM (2013) The nine-banded armadillo: a Venezuela. Mamm Biol 72:321–329 natural history. University of Oklahoma Press, Oklahoma Wetzel R (1985) and distribution of armadillos, Dasypodidae. McDonough CM, Loughry WJ (1997) Influences on activity patterns in a In: Montgomery GG (ed) The evolution and ecology of armadillos, population of nine-banded armadillos. J Mammal 78:932–941 sloths and vermilinguas. Smithsonian Institution Press, Washington, McDonough CM, Loughry WJ (2003) Armadillos (Dasypodidae). In: pp 23–46 Hutchins M (ed) Grzimek’s animal life encyclopedia, vol 13. Gale Worton BJ (1989) Kernel methods for estimating the utilization distribu- Group, Farmington Hills, pp 181–192 tion in home-range studies. Ecology 70:164–168